opencv的resize和matlab的imresize函數的計算

在用c++代碼復現matlab代碼時，遇到兩者resize函數的結果不相同的問題。

opencv:

resize(image1, reTmp, Size(50, 50), 0, 0, cv::INTER_LINEAR);

matlab:

tmp=imresize(img,[50 50],'bilinear','AntiAliasing',false);

以一張500×396的彩色圖片為例，resize到(50,50)：

 

matlab  imresize代碼：

img=imread('image/cluo.jpg');
img=single(img);
for i=1:10
    fprintf('%.6f ',img(1,i,1));
end
fprintf('\n');
tmp=imresize(img,[50 50],'bilinear','AntiAliasing',false);
for i=1:10
    fprintf('%.6f ',tmp(1,i,1));
end

c++ resize代碼：

#include <opencv2\opencv.hpp>
#include <iostream>
using namespace cv;
using namespace std;

void main()
{
    Mat image = imread("cluo.jpg");

    Mat image1(image.rows, image.cols, CV_32FC3);
    for (int i = 0; i < image.rows; i++)
        for (int j = 0; j < image.cols; j++)
        {
            image1.at<Vec3f>(i, j)[2] = image.at<Vec3b>(i, j)[0];
            image1.at<Vec3f>(i, j)[1] = image.at<Vec3b>(i, j)[1];
            image1.at<Vec3f>(i, j)[0] = image.at<Vec3b>(i, j)[2];
        }

    //打印輸出resize前r通道前10個像素點的坐標
    for (int i = 0; i < 10; i++)
        printf("%.6f ", image1.at<Vec3f>(0, i)[0]);
    printf("\n");
    Mat reTmp;
    resize(image1, reTmp, Size(50, 50), 0, 0, cv::INTER_LINEAR);
    //resize(image1, reTmp, Size(50, 50), 0, 0, cv::INTER_NEAREST);
    //打印輸出resize后r通道前10個像素點的坐標
    for (int i = 0; i < 10; i++)
        printf("%f ", reTmp.at<Vec3f>(0, i)[0]);
    printf("\n");

    system("pause");
}

結果比較：

結果基本相同

matlab中默認的     tmp=imresize(img,[50 50],'bilinear');    函數是打開了抗鋸齒   結果和opencv中的INTER_LINEAR不相同

 

matlab  和  opencv中  nearest  方式的結果比較：

c++:    resize(image1, reTmp, Size(50, 50), 0, 0, cv::INTER_NEAREST);

matlab:   tmp=imresize(img,[50 50],'nearest');

matlab:   tmp=imresize(img,[50 50],'nearest','AntiAliasing',false);

結果不相同